AI CONSULTANCY · FOR SMBs IN MENA · BASED IN DUBAI

We put AI to work where it makes a real difference.

A free diagnostic shows you where AI fits, what it would cost, and what to build first. Then we build it. Fixed-fee, scoped to your business. No retainers. No jargon. No projects that drag on.

Already know you need help? Book a 20-min call

WHAT MAKES US DIFFERENT
01

We measure before we build.

Most AI consultancies sell six-month strategy decks. Freelance shops sell builds before the build should happen. We do the measurement first, and we only build what the measurement justifies. The free diagnostic below is the entry point.

02

We publish what we won’t build.

Other consultancies hide their limits to keep options open in the sales cycle. We publish ours, because what we refuse tells you more about how we think than any list of services would.

  • Autonomous sales agents that close deals. Closing requires accountability. AI can prep and follow up. It shouldn’t commit your firm to anything.
  • Custom fine-tuned models for SMBs. Fine-tuning rarely beats RAG plus good prompts at your scale, and costs 5-10× more to maintain. We’ve seen this fail enough times to refuse to sell it.
  • Customer-facing legal, medical, or financial advice. The liability exposure of bad AI output in regulated advice exceeds anything an SMB can absorb. We build tools for licensed professionals, not advice for their clients.
  • “AI to replace your team.” Hiring decisions are management’s call, not ours. We sell augmentation that lets the same team do more, which is most of our catalog.
03

We ship with instrumentation, or we don’t ship.

Every workflow we build comes with the measurement to know whether it’s working. If we can’t define what success looks like in numbers, we don’t build it. This is also written into every Build contract as a kill-criterion: the metric that would tell us to shut a workflow down.

04

We’re built for the region we work in.

Sonata is headquartered in Dubai because the MENA SMB market is underserved relative to its appetite for serious AI work. We understand the local realities that generic AI consultancies miss: family-business decision structures, multi-entity group operations, regional data-residency expectations, the rhythms of GCC and Levant operating calendars, the demand for Arabic-language work alongside English. We work globally where the fit is right, but the centre of our practice is regional by design.

01 · DIAGNOSE · PREVIEW

A preview of your readiness score.

Eight quick questions, scored locally for a feel of the format. The real adaptive diagnostic at /diagnose asks questions tuned to your business, takes about ten minutes, and emails you a written report.
SONATA.DIAGNOSTIC v1.4 · SAMPLE
QUESTION 01 OF 08
Tell us, in one sentence. What does your business actually do?
STEP 01 / 08
~ 3 MIN
WHAT WE DO

Three stages of work, in order. Free, then paid, then production.

Stage 01 is the diagnostic above. Stages 02 and 03 are how it gets from a score to working software.
01 · DIAGNOSE

The Readiness Score

A 20-minute conversational diagnostic that probes four pillars (data, process, people, strategy) and returns a calibrated score plus a written report. You keep the report whether you hire us or not.

FORMATCONVERSATIONAL
WHENWEEK 0
PRICEFREE
02 · PRESCRIBE

The Application Map

Two weeks of discovery and synthesis. Output: a 12-20 page document ranking every viable AI application in your business by payback period, build complexity, and the failure modes we’ve seen elsewhere.

FORMATWORKSHOPS + REPORT
WHENWEEKS 1-2
FROM$8,000
03 · BUILD

The Implementation

Production builds of the workflows the map prescribed, with instrumentation from day one. Every workflow ships with a kill-criterion: the metric that would tell us to shut it down. You own the code, the data, the result.

FORMATFIXED-FEE PROJECT
WHENWEEKS 3-14
FROM$24,000
02 · THE READINESS MODEL

The four pillars we actually measure.

PILLAR 01Data. Where it lives, who owns it, how clean it is, whether anything connects to anything else.
PILLAR 02Process. What’s repeatable, what’s tribal knowledge, what already breaks under load.
PILLAR 03People. Skill, appetite, and where leadership is on the resistance-to-evangelism spectrum.
PILLAR 04Strategy. Whether you have a thesis on AI, or a vague feeling that you should.
SCORING TIERS
0-24Unmeasured
25-49Aware
50-74Calibrated
75-100Production-ready
03 · FIELD NOTES

How we’ve been thinking about AI lately.

Short essays on what works, what doesn’t, and what we’ve been arguing about internally. Updated when we have something worth saying.

Read all field notes

04 · QUESTIONS

What people ask us, before they hire us.

“You’ll be left behind.” Isn’t that what every AI pitch says?+
Yes, and it’s lazy. We don’t do fear-of-missing-out. We do arithmetic: your competitors who’ve deployed AI well are now operating at a measurably lower cost-per-customer. That gap compounds every quarter you don’t close it. The diagnostic exists to show you exactly how much that gap is worth in your business, not in someone else’s.
Why a score? Aren’t these things always too vague to measure?+
Vague when measured vaguely, yes. We use a calibrated rubric. Every score traces to specific signals in your data hygiene, process maturity, team composition, and stated strategy. The number is shorthand for a written report.
We’re a 30-person company. Isn’t AI overkill?+
The opposite, in our experience. SMBs are where AI moves the needle fastest, because one well-built workflow replaces a meaningful share of the manual work. We turn down enterprises with too many stakeholders more often than we turn down SMBs.
What if our score is low?+
Then we tell you what to fix before spending a dollar on AI. Sometimes the answer is “clean up your CRM and come back in six months.” That’s a free, honest report, not a sales call.
Are you tied to a particular AI vendor?+
No. We pick the model and stack that fit the job. We’ve shipped on Claude, GPT, open-source models, and combinations of all three. Fit-for-purpose, not brand-for-brand.
Who owns the work you build?+
You do. Full code, full data, full IP. We’re a consultancy, not a SaaS pretending to be one.
READY TO TALK?

Twenty minutes is all it takes to find out if we’re a fit.

No pitch, no slide deck. We’ll ask about your operation and tell you straight whether AI would actually help, where, and what it would cost. If we’re not the right team, we’ll say so.